@Gadgetoid I think this would be really useful. My only request would be if it could still hit a breakpoint if the debugger is attached (it looks like it'll still do this, but just checking.)

  while (1) {
        mp_event_handle_nowait();
        sleep_ms(1);
    }

Not sure about this detail, as if there's a full panic it's probably not safe to execute any event callbacks. EDIT: Although I realise without this TinyUSB may not ever deliver the panic message to the host... ah...

Is there any precedent for this sort of thing in other ports I could follow?

I think ports are pretty inconsistent in how they handle both hard faults and panics at the moment, a lot will sit in an infinite loop of some kind. Printing something and sitting still is pretty useful for development debug (especially if the USB serial port vanishes on reset). However, it's really inconvenient for production use if you've got a device deployed somewhere and you'd actually like it to reset itself. There are board hooks to add custom panic handlers on some ports, but they're not available on every port.

I have my own set of ideas about standardising this, but I haven't actually done anything about it. Something like what you've proposed here looks like a step in the right direction, to me.